LMI-based Synthesis of Anti-windup Filters with Pole Constraints

Authors:	Wang Chun-Chih, University of California at Berkeley, United States Tomizuka Masayoshi, University of California at Berkeley, United States
Topic:	2.1 Control Design
Session:	Nonlinear Control Systems II
Keywords:	Saturation control, Stability analysis, Control system synthesis, Constrained poles, Convex optimization.

Abstract

This paper presents a design method for a linear anti-windup filter with filter pole constraints. The anti-windup design problem can be cast as an LMI optimization problem. The anti-windup filter is optimized for $\mathcal{L}_2$ performance while its poles are constrained in a predefined region. The constrained poles ensure the fast recovery of saturation free linear responses and practical implementation of the anti-windup filter in digital form. This design method guarantees global stability for asymptotically stable plants, and local stability for plants, which are not asymptotically stable. The effectiveness of the design method is demonstrated by a simulation example.